1 September 2011Re/insurance

Challenges in the energy market

Following the Deepwater Horizon oil rig explosion on April 20, and subsequent extensive operations to clean up the spill, there has been considerable focus on what the future may hold for re/insurers covering structures in the Atlantic, US and Gulf of Mexico (GOM). Did the event have a sizeable enough re/insured component to produce marked changes in pricing in the energy and marine markets?

Aside from Deepwater, in 2010, insurers covering these regions have experienced modest softening of rates in hull, cargo and liability classes, but a disciplined market has generally been maintained.

The reinsurance market yields a similar picture—that of softening rates, but again, in a disciplined environment—and at the important January 1 reinsurance renewals period, hull, cargo and liability rates declined by around 10 percent compared to the same period last year, with certain territories experiencing larger decreases. In both non- catastrophe and pure marine reinsurance business, capacity is widely available, even in wind-exposed territories.

There is a similar scenario in the international reinsurance market, where capacity for energy risks is again abundant, and where non-GOM rates had decreased by around 10 to 15 percent. Undoubtedly, however, the recent events in the Gulf will have a marked impact on this trend.

Product evolution

Steps have been taken over the past decade to hone GOM marine and energy products, and poor profits performance has led insurers to try to structure solutions that achieve sufficient premium and granularity. After Hurricanes Katrina, Rita and Wilma, a strenuous effort was made to change the reinsurance product, because the process simply wasn’t working—both for cat losses and for run-of-the-mill losses—and even following Ike, reinsurers were still liable for a huge amount of the loss.

While retention levels have been considerably higher in recent years, much of the wind-exposure business is still transferred into the reinsurance market. Since Ike, retentions have become much larger and coverage has seen the application of defined wind limits, restrictions in respect of business insurance (BI) and contingent BI, the redrilling and plugging/abandonment of wells, and the scheduling of all assets covered by the original insurance policies.

Concerns over GOM pricing resulted in Lloyd’s initiating a review of its GOM exposures across all Syndicates. The conclusion may be that the energy market is still unsure of its wind exposures in the GOM. It is a real challenge for reinsurers to make their product a meaningful purchase for the buyer in the GOM market at the same time as making it a suitable assumed risk for the writer.

There is little doubt that the reinsurance community is generally content with the current product offered by insurers for wind coverage in the GOM. Assets are now scheduled and quantified, with appropriate limits and deductibles applied. Broad-brush coverages have been eliminated, resulting in a transparent product for both buyers and sellers. Accordingly, those reinsurers wishing to assume GOM wind exposure have a far better handle on what they are taking on and how much they can lose in any given loss scenario.

This transparency in itself generates a potential issue with the oil companies looking to buy insurance cover. Risk managers themselves are able to evaluate the risk covered by insurance and then determine if the product represents reasonable and/or meaningful value. Many of the larger operators, which historically have enjoyed large limits of indemnity and broad coverage, are far less enamoured of the current insurance offering. Some, particularly companies whose balance sheets dwarf those of the insurers they historically purchased from, have decided to retain their risks completely—either by not buying at all or by self-insuring via captives.

For the smaller to medium-sized companies, the issues are more complex. Their desire to purchase remains strong for a myriad of reasons, but the product in its current guise is criticised as offering little differentiation between buyers. Terms and conditions, particularly with regard to price, are often seen as a ‘tariff’ and do not sufficiently differentiate between what can be markedly different portfolios of assets.

Meanwhile, the liability market now has much the same set of issues to consider in the GOM, and perhaps globally, following the Deepwater oil spill. Insurance buyers have been able to purchase around $1 billion of protection at what appeared to be reasonable terms, while larger owners/operators were self-insured. However, new questions are being asked: what limits are available and what limits will be ‘required’ in future? What are the exposures and are they insurable? What are the right coverages, and terms and conditions?

Aside from the obvious dilemmas surrounding contractual liabilities, joint ventures, aggregations, contingent liabilities and a host of as yet unanswered questions, in order to provide insurance or reinsurance, there will have to be enough premium available to contemplate the potential for another Deepwater disaster. However, as yet, it does not appear that we are close to that eventuality.

The challenge for the re/insurance market and the associated modelling communities is to build on the work already conducted and to deliver a sustainable product that deals with these issues. Original assureds need the ability to purchase economically viable coverage that offers fair value for the risks being assumed by re/insurers. Credible and considered modelling assessments of these risks over the short, medium and long term will be instrumental in the success of viable re/insurance in this most testing of sectors.

Modelling the market

The modelling world has come a long way from the nascent suite of tools that came to market in the early 1990s, but it has not developed at a uniform rate across all classes of business and all global territories. In this regard, the modelling of marine and offshore energy risks has fallen behind and, in particular, the number of viable offshore models is still relatively limited.

Certain commercial model vendors have considerable maritime experience within their company history, and this has enabled them to build offshore models with wide-ranging functionality that are better able to capture and quantify re/insurers’ exposures. These models account for the nuances and knock-on effects in maritime losses—a key one being contingent BI. Contingent BI can have a huge impact on the resultant loss figure, but it is often overlooked in model development, with many vendors choosing to focus only on standard business interruption scenarios.

In short, the models that fall below par do so because they cannot capture the inherent complexity of insurance losses for offshore structures. This sizeable ‘miss factor’ has led Lloyd’s to undertake a review of the current suite of models in the market.

The historical inaccuracy of offshore models has not only dented the reputation of model vendors, but has also called into question the ability of re/insurers to accurately price offshore risk. Large losses in the GOM in recent years have raised rates for wind-related offshore cat. Large corporations, including many of the oil majors, are now retaining a significant portion of their risk (in some cases, they are fully retaining the risk and self-insuring through a captive structure).

There is an irony here, in that before catastrophe modelling became a significant component of the global re/insurance industry, re/insurers found it difficult to quantify and price for offshore cover, yet coverage was still being purchased by insureds. Modelling of offshore risks gave re/insurers their benchmark and a level of comfort in the rates they were charging, but the price point was considered too high by the corporations that required the cover, and so premium volumes began to decline.

Pricing is but one issue. Another is the ‘miss factor’ in the models’ quantification and qualification (or lack of it) of cargo and cargo losses. Typically, models do not account for cargo type, which can vary as a vessel travels between ports.

In short, there is a lack of refinement in the models, which leaves marine underwriters worrying about a potential accumulation of risk across their cargo books. No one has a grasp of the extent of these potential exposures, and so it is difficult to know whether the worries are justified at this point. However, there is no question that from a pricing perspective, the more information you have to hand, the better you can assess the risk and the better you can charge for cover. Achieving a more comprehensive cargo functionality will require the co-operation of the shipping industry, as well as investment by the model vendors.

While marine modelling has been found wanting, property risks are far better serviced, with models that are far more robust than those for offshore structures. This is mainly due to availability of data, data quality and loss experience. One of the reasons why property models have made enormous advances in the past decade is that they have been put to the test on many occasions and, while many would agree that they still have their faults, the loss experience has encouraged subsequent refinements, with each model upgrade being slightly more robust than its predecessor.

For offshore energy modelling, the same cannot be said. There have been a few tests of the models’ ability, but not enough to give the re/insurance industry a high level of confidence in their output.

In 2008, Hurricane Ike, the third-costliest hurricane ever to make landfall in the US, fanned the flames of inherent scepticism surrounding offshore models, with losses that were above and beyond modelled forecasts. Ike highlighted that the models needed more testing and that they hadn’t reached the level of sophistication of the property models.

The next generation of software is now coming to market, including ImpactOnDemand from Aon Benfield, which will allow re/insurers to track weather events in real time, thanks to frequent data streams from Tropical Storm Risk. This will help them to plot the course of storms and to elucidate the potential impact of weather events on their portfolios.

An active season?

Tropical Storm Risk (TSR), a meteorological group affiliated to Aon Benfield, believes that the 2010 Atlantic hurricane season will be 60 percent more active than the long-term average. TSR’s resident experts, Professor Mark Saunders and Dr. Adam Lea, expect that around 16 named storms, nine hurricanes and four major hurricanes (defined as Category 3+) will occur in the Atlantic region this year.

Furthermore, Saunders and Lea believe that this hurricane season has a high probability of being in the top third of all hurricane seasons in terms of activity, due to forecast trade winds in the upper levels of the atmosphere during August and September, as well as expected above- average sea surface temperatures.

Meanwhile, the longer-term outlook for hurricane activity in light of climate change is still under debate. According to Robert Pielke Jr, a professor at the University of Colorado’s Environmental Studies Programme and former director of the University’s Center for Science and Technology Policy Research, while the costs of catastrophic weather are escalating rapidly, there is no proof of a link between these rising costs and global warming. After adjusting for increases in asset volumes and values, catastrophic events are costing global communities no more today than they did 50 years ago, which presents a conundrum for re/insurers from a pricing perspective as to whether they can account for climate change in their calculations.

What is predictable is that if the models themselves were to become more reliable, a convergence in pricing between what re/insurers think they should charge and what buyers think they should pay would no doubt be witnessed.

Graeme Moore is chief executive officer of Aon Benfield Global Re Specialty. He can be contacted at:

Paul Markey is chairman of Aon Benfield Bermuda. He can be contacted at: